Abstract:

With the continuous development and extension of 95598 business,the intensity of manual telephone traffic increases.In order to further deepen the consciousness and understanding of the hidden features and the demands of the customers,improve the customer service level of 95598,the typical application scenarios in the customer service such as the tendency of the complaint were refined.Based on the data of power service orders,the key index of modeling was selected.Through entropy weight method,principal component analysis and decision tree and other data mining algorithms,potential complaint propensity customers and planned blackout sensitive customers in order to carry out targeted service resource scheduling were identified,fully do a good job of response measures,effectively reduce complaint pressure and improve service accuracy.

Key words: complaint work order, data mining, entropy weight method, decision tree